J. Embryol. exp. Morph. 80, 63-74 (1984)
Printed in Great Britain (£) The Company of Biologists Limited 1984
The mode of cell death associated with cavitation in
teratocarcinoma-derived embryoid bodies
By SHEILAGH M. BOYD, M. L. HOOPER AND A. H.
WYLLIE
Department of Pathology, University of Edinburgh Medical School, Teviot
Place, Edinburgh, EH8 9AG, U.K.
SUMMARY
Cell death occurring in embryoid bodies derived from the embryonal carcinoma cell line,
PSA4, which undergo cavitation, and in those from the related cell line S2, which do not
undergo cavitation, was classified as apoptosis or necrosis by ultrastructural criteria. Both
modes of cell death were seen in PS A4 embryoid bodies while apoptosis alone was seen in S2
embryoid bodies. No significant difference was found between PSA4 and S2 embryoid bodies
either in apoptotic incidence score or in the spatial distribution of apoptotic events. We
therefore conclude that although apoptosis and tissue modelling coexist in PSA4 embryoid
bodies, necrosis rather than apoptosis is causally related to formation of the cavity.
INTRODUCTION
The significance of cell death in early development has long been recognized
(Gliicksmann, 1951; Bellairs, 1961; Saunders, 1966). Study of its mechanism is
hindered however by the small size of the early embryo, the admixture of the
dying cells with their proliferating and differentiating neighbours, and the
limited range of experimental manipulation which can be applied to early embryonic tissues. Cell lines derived from the pluripotent embryonal carcinoma
(EC) cells of teratocarcinomas afford more plentiful and accessible material than
embryonic tissues, and have permitted elucidation of several basic questions in
developmental biology (Martin, 1980). When cells of certain embryonal carcinoma lines are grown in suspension culture, two-layered cell aggregates are
formed in which a core of EC cells is surrounded by an outer layer of endoderm.
In these 'embryoid bodies' reproducible and focal cell death associated with
cavity formation has been described (Martin, Wiley & Damjanov, 1977). As EC
cells are thought to be developmentally analogous to epiblast cells of the early
post-implantation embryo, it has been suggested that the cavity formation in the
embryoid bodies may parallel the formation of the pro-amniotic cavity. This
suggestion is supported by the observation that further differentiated cell types
appear in the embryoid body concurrently with, or subsequent to, cavity formation. EC cell lines exist whose embryoid bodies neither develop cavities nor
differentiate beyond the two-layered stage in suspension. Comparison of related
64
S. M. BOYD, M. L. HOOPER AND A. H. WYLLIE
cavitating and non-cavitating cell lines should therefore allow the degree of
association between cell death and cavity formation to be precisely assessed.
Although there have been many observations on cell death in developing
systems, only recently has the classification of death been formalized to emphasize the existence of two, apparently discrete, morphologically defined pathways - necrosis and apoptosis (Kerr, Wyllie & Currie, 1972; Wyllie, 1981).
Whereas necrosis is characterized by cellular swelling and rupture of the plasma
and internal membranes, apoptosis is characterized by cell shrinkage, with conservation of the integrity of cytoplasmic organelles, but widespread condensation of nuclear chromatin, and cleavage of chromatin into oligonucleosome
chains (Wyllie, 1980). Extensive review of the literature suggests that apoptosis
rather than necrosis is associated with developmentally regulated death (Wyllie,
Kerr & Currie, 1980). Only in phylogenetically primitive organisms has necrosis
alone been described unequivocally as a homeostatically regulated mode of cell
death (Bowen & Ryder, 1974). The cell death associated with cavity formation
in embryoid bodies has been interpreted as apoptosis (Solter & Damjanov,
1979). In this paper we present data which do not support this contention. We
describe and quantify cell death in embryoid bodies of the EC cell line PSA4,
which form cavities, and in those of the related EC line S2, which do not. We
show that whilst apoptosis occurs in both cases, necrosis appears exclusively in
the PSA4 embryoid bodies, in association with cavity formation.
MATERIALS AND METHODS
Cell culture
PSA4 and S2 EC cell lines are derived respectively from primary and secondary cultures of the transplantable tumour line OTT5568 (Martin & Evans,
1975a). PSA4TG12 is a thioguanine-resistant variant of PSA4 which behaves
identically to its parent with regard to cavitation and differentiation (Slack,
Morgan & Hooper, 1977). Stock cultures of these pluripotent EC cell lines were
maintained in their undifferentiated state by frequent subculture on a feeder
layer of mitomycin C-treated mouse STO fibroblasts (Martin & Evans, 19756),
in Glasgow-modified Eagle's medium supplemented by non-essential amino
acids (each at 0-lmM), 1 mM-sodium pyruvate, 0-1 mM-/3-mercapto-ethanol, and
10 % newborn calf serum (selected batches). Cultures were maintained at 37 °C
in a humidified atmosphere of 95 % air, 5 % CO2. Embryoid body formation was
induced as described by Martin & Evans (19756). Cells were seeded at 2-5 x 106 per
5 cm diameter tissue culture dish in medium as above but lacking non-essential
amino acids, pyruvate and /3-mercapto-ethanol. On the fifth day after seeding,
nests of cells formed and were gently detached by blowing fresh medium across
the surface. These aggregates were subsequently grown in suspension in bacteriological dishes.
Cell death in cavitating embryoid bodies
65
Morphological studies
Embryoid bodies on the day of detachment and 1, 3, 5, 7 and 9 days
subsequently were transferred to universal containers and allowed to sediment
under gravity. They were then washed twice in Dulbecco's phosphate-buffered
saline solution A (PBS A) and fixed in Bouin's fluid. 5 jLtm paraffin sections were
stained with Mayer's haematoxylin and eosin or by the Feulgen reaction.
Feulgen-stained preparations were scanned at 250 x magnification using a Tektronic (4050 series) Graphic Information System to determine the following
parameters in each embryoid body section: (1) area; (2) area of cavity; (3)
number of apoptotic cells. From these data a score proportional to the incidence
of apoptosis could be derived. This 'apoptotic incidence score' was defined as the
number of apoptotic cells per unit area of embryoid body section. It is recognized
that the absolute incidence of apoptosis in embryoid bodies is a complicated
function of this simple score, requiring knowledge of the duration for which an
apoptotic body, once formed, remains identifiable by light microscopy. Further,
an absolute score of incidence requires information about the effective diameter
of apoptotic cells relative to the section thickness, as this affects the probability
with which they are recognized (Aherne, 1967). However, since these additional
factors are likely to be constant in embryoid bodies, the simple score defined
above permits analysis of changes in incidence of apoptosis.
For electron microscopy, embryoid bodies were fixed in 3.1 % glutaraldehyde
and 0-lM-osmium tetroxide, embedded in Araldite and stained with uranyl
acetate and lead citrate. Thin sections were viewed in a Jeol 100/S electron
microscope.
RESULTS
As described by Martin et al. (1977) PSA4 EC cells formed embryoid bodies
with eccentric cavities. The cavities were first observed on days 4-5 after detachment, as areas of cellular debris. They were lined by cells which had a radial
polarity, contrasting with the apparently random orientation of the EC cells in
the rest of the embryoid body core. Over the next 4-5 days the cavities enlarged
(Figs 1, 2) and differentiated cells appeared elsewhere in the embryoid bodies.
S2 cells formed simple embryoid bodies in which neither cavity formation nor
morphological evidence of differentiation beyond the two-layered stage was
observed.
Two types of cell death were identifiable in embryoid bodies. The first was
observed in both PSA4 (cavitating) and S2 (non-cavitating) embryoid bodies and
bore the characteristics of apoptosis. Light microscopy showed that the affected
cells were scattered amongst apparently viable neighbours (Fig. 3). Their nuclei
were rounded, sometimes fragmented, and stained homogeneously in the Feulgen reaction. Ultrastructurally some of these cells appeared as membrane-
S. M. BOYD, M. L. HOOPER AND A. H. WYLLIE
B
Figl
Cell death in cavitating embryoid bodies
67
30000-
25000-
g
1
20000-
-5
-4
<>
/
- 3
15 000"
-2
10000-
-1
,-_-
^
Q
D
Q
-0
5000.
Time in suspension (days)
Fig. 2. Time course of embryoid body growth and cavitation in PSA4 and S2 embryoid bodies, showing cross-sectional area (left ordinate) of PSA4 (•) and S2 (O)
embryoid bodies, and percentage of embryoid body cross-sectional area occupied by
cavity (right ordinate) for PSA4 (•) and S2 (•). Data are presented as the mean
determined from five separate experiments, with the error bar indicating one standard error of the mean.
bounded fragments including compacted cytoplasmic organelles and nuclei with
condensed chromatin (Fig. 4B). In other cells, presumably at an earlier stage, the
condensation of nuclear chromatin was localized to a hemilunar cap (Fig. 4C).
Although occasionally observed free in the intercellular space (Fig. 4B), apoptotic bodies were more often found undergoing degradation within large
heterophagosomes in viable EC cells. Apoptotic incidence scores were closely
similar in PSA4 and S2 embryoid bodies (Fig. 5). Around the onset of cavity
formation in PSA4 embryoid bodies there was no increase in apoptotic incidence. In both types of embryoid body the apoptotic cells showed similar
spatial distributions (Fig. 3) with no focal grouping in the vicinity of developing
cavities.
Fig. 1. Embryoid bodies from (A) PSA4TG12, at day 5 after detachment; (B) S2,
at 7 days after detachment (both panels xl44; scale bar, 100/urn).
68
S. M. BOYD, M. L. HOOPER AND A. H. WYLLIE
PI
S2
S4
Fig. 3. Spatial distribution of apoptotic cells within embryoid bodies of PSA4 and
S2. Outlines of sections taken at approx. 20/im intervals from serial sections of a
PSA4 embryoid body at day 3 after detachment (P1-P4), an S2 embryoid body at day
3 (SI, S2) and an S2 embryoid body at day 4 (S3, S4). Apoptotic cells are shown by
dark circles ( • ) . In sections P3 and P4 a developing cavity is present, within which
are necrotic cells (stippled areas). x!35; scale bar, 100jum.
Cell death in cavitating embryoid bodies
69
The second type of cell death was observed only in PSA4 embryoid bodies, and
showed the features of necrosis. At light microscope resolution the dead cells
were represented by debris in the developing cavities. The earliest recognizable
cavities were often completely filled with these necrotic cells. Ultrastructurally,
nuclei were swollen and included heterochromatin aggregates in approximately
the same distribution as in normal nuclei (Fig. 4D). Cytoplasmic swelling was
gross and frequently associated with ruptured plasma membranes, 'high amplitude' swelling of mitochondria, or extensive disruption of ER membranes
(Figs4D,E).
DISCUSSION
The results show conclusively that apoptosis alone is not responsible for cavity
formation in cystic embryoid bodies. Although all the characteristic features of
apoptosis were identified in embryoid bodies during culture, the spatial and
temporal distribution of apoptotic cells within the embryoid bodies was unrelated to the appearance of cavities. Moreover apoptosis was as frequent in noncavitating S2 as in cavitating PSA4 embryoid bodies. Cell death, however, did
occur during the development of the cavities but with the morphology of
necrosis. In contrast to apoptotic cells, the necrotic cells showed focal grouping
and at the onset of cavity formation the entire region of the developing cavity was
frequently packed with necrotic cells. The earliest morphological stigmata we
could identify - nuclear and cytoplasmic oedema - excluded the possibility that
these dying cells were undergoing degeneration secondary to an initial phase of
apoptosis as has been described elsewhere (Wyllie et al. 1980). This result was
unexpected as, unlike apoptosis, necrosis is in general associated with pathological cell states, often involving grossly unphysiological environmental conditions
(Kerr et al. 1972; Wyllie, 1981).
Cell death has been described in the early mouse embryo at the blastocyst
stage (El-Shershaby & Hinchliffe, 1974; Copp, 1978), in the egg cylinder (Bonnevie, 1950; Poelmann & Vermeij-Keers, 1976), and in association with neural
tube formation (Schluter, 1973; Poelmann, 1980). Bonnevie (1950) stated that
cell death occurred at the inner surface of extraembryonic and embryonic
ectoderm, beginning after a proamniotic cavity was already evident. In contrast
Poelmann & Vermeij-Keers (1976) observed cell death (with the morphology of
apoptosis) at about the time of proamniotic cavity formation but with no obvious
clustering at the prospective site of cavity formation. Despite these differences
in detail, in neither case was cell death associated spatially and temporally with
proamniotic cavity formation, nor is there evidence of this in other published
studies of the egg cylinder at the light-microscope level (Reinius, 1965; Theiler,
1972; Poelmann, 1975). In contrast, in embryoid bodies the prospective cavity
is first recognizable as an area of dying cells (Martin et al. 1977; Fig. 3). While
this does not prove that the necrosis is the primary event in cavitation (we cannot
70
S. M. BOYD, M. L. HOOPER AND A. H. WYLLIE
exclude the possibility that necrosis and cavitation coexist as parallel results of
a common stimulus) it does suggest that embryoid body cavitation may not
provide a model which can be applied in detail to the specific case of proamniotic
cavity formation in the embryo. Although Martin et al. (1977) found that the
development of embryoid bodies in vitro paralleled normal embryonic development in a general way, they, too, pointed out discrepancies in detail, notably in
endoderm and mesoderm formation, and noted that embryoid body development
Fig. 4
Cell death in cavitating embryoid bodies
l-4-i
1-2-
2
i-o-
x
0-8"
0-6-
0-4-
3
5
7
Time in suspension (days)
Fig. 5. Time course of apoptotic incidence in PSA4 (•) and S2 (O) embryoid bodies.
Data are presented as the mean determined from five separate experiments, with the
error bar indicating one standard error of the mean.
Fig. 4. Ultrastructure of normal, apoptotic and necrotic cells in embryoid bodies.
(A) Normal EC cell from an S2 embryoid body. Note the dispersed nuclear
chromatin, prominent nucleoli and cytoplasm rich in polysomes (x3800).
(B) Fragments of an apoptotic EC cell lie in the intercellular space in a PSA4
embryoid body. One fragment shows the characteristic compaction of cytoplasmic
organelles (a) whilst another consists almost entirely of condensed chromatin (b)
(X3800).
(C) Higher magnification of an apoptotic EC cell from a PSA4 embryoid body.
The smooth plasma membrane, cytoplasmic electron density, vesicular ER, and
nucleus with prominent cap of condensed chromatin are characteristic features. In
contrast to the surface of the apoptotic cell, that of adjacent normal cells shows
microvilli and active endocytosis (x9600).
(D) A necrotic EC cell from a cavitating PSA4 embryoid body. Mitochondria, in
contrast to those in Fig. 4B, show disrupted cristae and prominent matrix densities
(arrow). The nuclear chromatin remains dispersed, and a residual nucleolar structure is present despite the advanced stage of degradation of this cell (x5100).
(E) Cellular debris, some without bounding membranes, from the necrotic centre
of a PSA4 embryoid body cavity (x3200).
In all panels, scale bar = 5 [im.
71
72
S. M. BOYD, M. L. HOOPER AND A. H. WYLLIE
more closely paralleled that of normal inner cell masses explanted in culture
(Wiley, Spindle & Pedersen, 1978). Embryoid body cavitation is nevertheless a
developmentally significant tissue-modelling process, since further differentiation of the core EC cells is restricted to those embryoid bodies which show
cavitation, and the first cells to differentiate are aligned along the edge of the
cavity.
Two facts argue against the hypothesis that necrosis in PSA4 embryoid bodies
is merely a trivial culture artefact, due, for example, to inadequate diffusion of
oxygen or substrate to the central cells in the embryoid body. Firstly, necrosis is
always absent in S2 embryoid bodies, even at diameters twice as great as those
found in PSA4 embryoid bodies with internal necrosis, and despite the similar
overall growth rates of cavitating and non-cavitating embryoid bodies. Secondly,
cavitation in PSA4 embryoid bodies is always eccentric, and frequently very
markedly so, whereas hypoxic or nutritional necrosis is centrally disposed in
'spheroids' of other cell types growing in suspension culture (Franko & Sutherland, 1979) or occurs at a remarkably constant, predictable distance from blood
vessels in tissues (Thomlinson & Gray, 1955; Tannock, 1970).
Two classes of hypothesis could account for the differing incidence of necrosis
in PSA4 and S2 embryoid bodies. (1) S2 EC cells may be incapable, perhaps by
virtue of intrinsic metabolic differences, of responding to lethal stimuli equally
present in PSA4 and S2 embryoid bodies. (2) Conversely, lethal stimuli may be
engendered only in PSA4, and not in S2 embryoid bodies. The latter situation
could arise, for example, because of differences in the nature and extent of
differentiation prior to cavitation. At least one such difference in differentiation
is recognized prior to cavitation: PSA4 embryoid bodies form both visceral and
parietal endoderm on their surfaces, whereas S2 embryoid bodies form only
parietal endoderm (Martin et al. 1977). The question remains open whether the
wider spectrum of differentiation which develops subsequently in PSA4 embryoid bodies is a consequence of cavitation, or whether the primary events in
differentiation precede cavitation. It may be possible to distinguish between
hypotheses (1) and (2) by study of mosaic embryoid bodies formed from mixtures
of cells from cavitating and non-cavitating lines; in our laboratory we are currently
developing the appropriate lineage markers to permit analysis of such mosaic
embryoid bodies.
In conclusion, the data presented here emphasize the need for caution in
interpreting the role of cell death in morphogenesis. Apoptosis and cavity formation are not causally related in embryoid bodies, although they coexist in the
PSA4 line. Necrosis, however, is closely associated with cavity formation, but we
have not established which is the primary event.
We are grateful to Robert Hogg, Stuart MacKenzie and Alan Smith for technical assistance,
to Tom Smith for helpful discussion, and to Alec McLean for help with computer programming for image analysis. M.L.H. and A.H.W. are supported by grants SP1600 and SP1370 and
A.H.W. by a Career Development Award from the Cancer Research Campaign.
Cell death in cavitating embryoid bodies
73
REFERENCES
W. (1967). Methods of counting discrete tissue components in microscopical sections. /. Roy. microsc. Soc. 87, 493-508.
BELLAIRS, R. (1961). Cell death in chick embryos as studied by electron microscopy. /. Anat.
95, 54-60.
BONNEVIE, K. (1950). New facts on mesoderm formation and proamnion derivatives in the
normal mouse embryo. /. Morph. 86, 495-505.
BOWEN, 1. D. & RYDER, T. A. (1974). Cell autolysis and deletion in the planarian Polycelis
tenuis Iijima. Cell Tissue Res. 169, 193-209.
COPP, A. (1978). Interaction between inner cell mass and trophectoderm of the mouse blastocyst: I. A study of cellular proliferation. J. Embryol. exp. Morph. 48, 109-125.
EL-SHERSHABY, A. M. & HINCHLIFFE, J. R. (1974). Cell redundancy in the zona-intact preimplantation mouse blastocyst: a light and electron microscope study of dead cells and their
fate. J. Embryol. exp. Morph. 31, 643-654.
FRANKO, A. J. & SUTHERLAND, R. M. (1979). Oxygen diffusion distance and development of
necrosis in multicell spheroids. Radiat. Res. 79, 439-453.
GLUCKSMANN, A. (1951). Cell death in normal vertebrate ontogeny. Biol. Rev. 26, 59-86.
KERR, J. F. R., WYLLIE, A. H. & CURRIE, A. R. (1972). Apoptosis: a basic biological
phenomenon with wide-ranging implications in tissue kinetics. Br. J. Cancer 26, 239257.
MARTIN, G. R. (1980). Teratocarcinomas and mammalian embryogenesis. Science 209,
768-775.
MARTIN, G. R. & EVANS, M. J. (1975a). The formation of embryoid bodies "in vitro" by
homogeneous embryonal carcinoma cell cultures derived from single cells. In Roche Symposium on Teratomas and Differentiation (eds M. I. Sherman & D. Solter), pp. 169-187.
New York: Academic Press.
MARTIN, G. R. & EVANS, M. J. (19756). Differentiation of clonal lines of teratocarcinoma
cells: formation of embryoid bodies in vitro. Proc. natn. Acad. Sci., U.S.A. 72,1441-1445.
MARTIN, G. R., WILEY, L. M. & DAMJANOV, I. (1977). The development of cystic embryoid
bodies in vitro from clonal teratocarcinoma stem cells. Devi Biol. 61, 230-244.
POELMANN, R. E. (1975). An ultrastructural study of implanting mouse blastocysts: coated
vesicles and epithelium formation. J. Anat. 119, 421-434.
POELMANN, R. E. (1980). Differential mitosis and degeneration patterns in relation to the
alterations in the shape of the embryonic ectoderm of early post-implantation mouse
embryos. J. Embryol. exp. Morph. 55, 33-51.
POELMANN, R. E. & VERMEIJ-KEERS, C. (1976). Cell degeneration in the mouse embryo: a
prerequisite for normal development. In Progress in Differentiation Research (ed. N.
Miiller-B6rat et al.), pp. 93-102. Amsterdam: N. Holland Publishing Company.
REINIUS, S. (1965). Morphology of the mouse embryo from the time of its implantation to
mesoderm formation. Z. Zellforsch. mikrosk. Anat. 68, 711-723.
SAUNDERS, J. W. (1966). Death in embryonic systems. Science 154, 604-612.
SCHLUTER, G. (1973). Ultrastructural observations on cell necrosis during formation of the
neural tube in mouse embryos. Z. Anat. EntwGesch. 141, 251-264.
SLACK, J. R., MORGAN, R. H. M. & HOOPER, M. L. (1977). Action potentials recorded
intracellularly from differentiated mouse teratocarcinoma cells in culture. Expl Cell Res.
107, 457-459.
SOLTER, D. & DAMJANOV, I. (1979). Teratocarcinoma and the expression of oncodevelopmental genes. In Methods in Cancer Research. Vol. 18 (eds W. H. Fishman & H. Busch),
pp. 277-332. New York: Academic Press.
TANNOCK, I. F. (1968). The relation between cell proliferation and the vascular system in a
transplanted mouse mammary tumour. Brit. J. Cancer 22, 258-273.
THEILER, K. (1972). The house mouse. Springer-Verlag, Berlin.
THOMLINSON, R. H. & GRAY, L. H. (1955). The histological structure of some human lung
cancers and the possible implications for radiotherapy. Brit. J. Cancer 9, 539-549.
AHERNE,
74
S. M. BOYD, M. L. HOOPER AND A. H. WYLLIE
L. M., SPINDLE, A. & PEDERSEN, R. (1978). Morphology of isolated mouse inner cell
masses developing in vitro. Devi Biol. 63, 1-10.
WYLLIE, A. H. (1980). Glucocorticoid-induced thymocyte apoptosis is associated with endogenous endonuclease activation. Nature (Lond.) 284, 555-556.
WYLLIE, A. H. (1981). Cell death: a new classification separating apoptosis from necrosis. In
Cell Death in Biology and Pathology (eds I. D. Bowen & R. A. Lockshin), pp. 9-34.
London and New York: Chapman and Hall.
WYLLIE, A. H., KERR, J. F. R. & CURRIE, A. R. (1980). Cell death: the significance of
apoptosis. Int. Rev. Cytol. 68, 251-306.
WILEY,
{Accepted 17 November 1983)